LPL 2021 Summer Playoffs Prediction

It's time for the LPL Summer Playoffs! We'll start by looking back over the season and our initial predictions, and then, as always, simulate the playoffs to get you the scoop on who we think will win and qualify to Worlds! (or qualify via Championship Points)

Pre-Season Predictions

At the start of the summer season, we had generated this power ranking of of the LPL teams, with our expected records for each. And once again, the same caveat as before applies: these predictions were all done before the season with the information we had at the time; throughout the season the model updated itself with information from new matches and was able to give us more up-to-date predictions (as well as us being able to keep track of roster changes), which feeds into our suggested bets.

So, what happened?

Quite a few upsets and subverted expectations, as it turns out. Very few teams performed how either we or the market would have expected them to. The higher-predicted teams generally did better, but there were a number of teams that significantly outperformed expectations.

Bilibili, LNG, and Oh My God all strongly outperformed. Rare Atom and Team WE also did a bit better than we thought they would, though the estimates weren't horribly off. In fact, this allowed Oh My God to just barely squeeze into the playoffs with a 9-7 record.

Victory Five, on the other hand, went home with a 0-16 record, which is... disappointing, to say the least. JD and Invictus also had somewhat lackluster seasons. JD's 7-9 record was two wins off of the cutoff for playoffs, so it's not like it was a one-match difference maker, but they really should have been a strong contender.

Bets

Over the course of this season, we found betting value on 46 of 136 matches - just about one in three. This is slightly higher than both our LEC and LCS bet rates (both closer to 25%), so we disagreed with the market a bit more often here.

Overall, we would be up $767.48 out of $7406.78 in total wagers (not all at the same time, though, so you wouldn't have needed that much money up front):

Our biggest single win was a $127 bet on Rogue Warriors vs. Rare Atom, paying out at 6.94. We thought they had roughly a 25% chance of an upset, compared to the market-implied 14%. Of course, on this one individual bet we were still likely to lose, but make enough bets like this and eventually they pay off! This time, a $752 payout was the reward.

This was a particularly upset-heavy season, as it turns out. Underdogs won way more than their expected share of games, either by our expectations or the market's. In fact, if you had simply bet on the underdog in each match, you'd have been profitable (normally, always betting on the underdog, or always on the favorite, should roughly even out if the market is placing lines well, and should steadily bleed money due to the vig):

Playoff Bracket

Enough living in the past, let's look to the future!

The playoffs feature a somewhat unique bracket style: first, there are two single-elimination, "king-of-the-hill"-style brackets seeded by summer season performance. Those then feed into a more standard double-elimination bracket. All matches are best of 5, as is typical for playoffs:

The overall winner gets a direct qualification to Worlds, while other places just get Championship Points which can indirectly lead them to Worlds as well. FunPlus Phoenix and EDward Gaming have a huge leg up by starting in the second phase of the bracket. For everyone else, obviously, the farther back you start the harder the journey is. Can't count anyone out though - there's a reason we still go through these full playoffs and don't just take the top couple of teams, and it's because anything can happen!

Simulating the Bracket

Running this bracket through 10000 iterations of a Monte Carlo simulation, here is the probability for each team to win overall, as well as their median placement:

As expected, FPX and EDG have a huge advantage here, by being the only two teams guaranteed to be in the double elimination portion of the bracket (and therefore guaranteed at least 4th place). RNG has a reasonable shot at first as well. Interestingly, TES is actually more likely to take 1st than RA in spite of starting back a bracket (due to their more favored later matchups), but will place worse on average.

Whereas in the LEC and LCS playoffs, top three teams qualify for worlds, the LPL works a bit differently: the top team places for Worlds, but after that, the team with the most championship points also qualifies! Points are awarded as such (AQ = Automatic Qualification):

These points are additive with spring season points (table limited to teams in the playoffs):

With that in mind, here's the probability of each team qualifying to Worlds, either by winning the playoffs or by having the most Championship Points:

FPX, EDG, and RNG are all more likely than not to at least qualify, though TES might make a showing as well. In our 10000 runs, OMG never once won first, or placed highly enough (and other teams placed poorly enough) to take home the most Championship Points.

The bottom six teams, unfortunately, have almost no shot at it. In fact, the odds of any one of them qualifying is 6.83%, which is never impossible but fairly unlikely.

It's not quite over for any teams that don't make it, however. After this event, two teams will qualify for the Worlds Main Event. However, China will still have regional qualifiers afterwards for the next four highest-placed teams in Championship Points, in another bracket in which first place qualifies for the Main Event and second place qualifies for the Play-In. We'll cover that when things get to that point, but it's worth noting that this isn't the end-all-be-all for these teams.

And, finally, here's the distribution of each team's outcomes:

Most teams' performances are concentrated on one side of the distribution, which makes sense given the cutthroat, king-of-the-hill style bracket. Even if you make it past your first match, you're still one loss away from a quick exit.

Favoritism

Here's how the final standings would look if the favorite won every single match:

This mostly lines up with our previous expectations of placement; the only big difference is that RNG places above EDG here - in each head-to-head on the way they're favored, including a roughly 60/40 edge against EDG itself, but because they start lower down they've got more chances to run into trouble and therefore have a lower overall chance of taking 1st.

As always, this bracket shouldn't be taken as gospel; upsets happen all the time (especially in the LPL, as evidenced by the summer season), and the odds of this exact bracket happening are only 0.39%.

(with 12 matches in the bracket, there are 2¹² = 4096 possible brackets, so the average bracket has a 1/4096 = 0.024% chance of occurring. This "favorites" bracket, therefore, is ~16 times more likely than the average.)

Bets Ahoy!

As with other tournaments, there doesn't seem to be a good way to bet outright on the winner of the tournament, or other standings. Each individual match will have its own odds though, so make sure you keep up with our suggested bets to keep up to date!

This analysis was mostly meant to illustrate possible ways this tournament could play out, with some numbers to back it up. With a 6.83% chance of some pretty major upsets, though, and LPL's tendency to produce upsets, it's going to be a fun one to watch. What curveball are you hoping for?

For more in-depth match analysis, discussion, and predictions, as well as quantitative betting strategy tips, subscribe to stay in the loop! You can also follow us on Twitter or Facebook for quick prediction snapshots of upcoming matches. Alacrity is your fastest path from simply gambling to beating the market. Stay ahead of the game!